omh-jit-learn
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided context regarding project blockers and knowledge levels to generate learning briefs. This ingestion of untrusted data represents a potential attack surface. However, the skill mitigates this by requiring a formal 'confirmed target statement' before conducting research and limiting the agent's output to a strictly defined Markdown structure.
- [COMMAND_EXECUTION]: The instructions include calls to the
omhcommand-line utility for runtime recording and state management (e.g.,omh runtime record). These appear to be standard instrumentation tools provided by the skill's author ('oh-my-hermes') for platform-specific workflow tracking and do not involve the execution of untrusted code or arbitrary user input.
Audit Metadata